estimtf

The estimtf package provides functions to find the Maximum Likelihood Estimates of parameters from probability distributions and linear regression models using the TensorFlow optimizers.

Installation

You can install estimtf from GitHub. It is recommended to follow these steps to avoid problems when using the package:

# Step 1: Install the reticulate package
install.packages("reticulate")
library(reticulate)

# Step 2: Install the tensorflow package
install.packages("tensorflow")
library(tensorflow)

# Step 3: Use the install_tensorflow() funcion to install the TensorFlow module
install_tensorflow()

# Step 4: Confirm that the TensorFlow installation succeded
library(tensorflow)
tf$constant("Hello Tensorflow")

# Step 5: Install the devtools package
install.packages("devtools")

# Step 6: Install and load the estimtf package
devtools::install_github("SaraGarcesCespedes/estimtf", force=TRUE)
library(estimtf)

Example

This is a basic example that shows how to estimate the mean and standard deviation parameters from the normal distribution using the mle_tf function:

# Load the estimtf package
library(estimtf)

# Estimation of parameters mean and sd from the normal distribution

# Generate a sample from the normal distribution
x <- rnorm(n = 1000, mean = 10, sd = 3)

# Find the MLE of the parameters using the mle_tf function
estimation <- mle_tf(x, 
                     xdist = "Normal", 
                     optimizer = "AdamOptimizer",
                     initparam = list(mean = 0.5, sd = 0.5),
                     hyperparameters = list(learning_rate = 0.1))

# Get the summary of the estimates
summary(estimation)
#> Distribution: Normal 
#> Number of observations: 1000 
#> TensorFlow optimizer: AdamOptimizer 
#> Negative log-likelihood: 668.5151 
#> Loss function convergence, 2071 iterations needed. 
#> ---------------------------------------------------
#>      Estimate  Std. Error Z value Pr(>|z|)    
#> mean  10.03652    0.09315   107.7   <2e-16 ***
#> sd     2.94563    0.06517    45.2   <2e-16 ***
#> ---
#> Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

You can visit the package website to explore the function reference. Also, to learn more about how to use the functions of the estimtf package, visit this Colab notebook that includes examples with multiple distributions and linear regression models.